Recycled dehydrated lithosphere observed in ... - Semantic Scholar
articles
Recycled dehydrated lithosphere observed in plume-influenced mid-ocean-ridge basalt Jacqueline Eaby Dixon*, Loretta Leist*, Charles Langmuir† & Jean-Guy Schilling‡ * RSMAS/MGG, University of Miami, 4600 Rickenbacker Causeway, Miami, Florida 33149, USA † Department of Earth & Planetary Sciences, Harvard University, Cambridge, Massachusetts 02138, USA ‡ Graduate School of Oceanography, University of Rhode Island, Narragansett, Rhode Island 02882, USA
...........................................................................................................................................................................................................................
A substantial uncertainty in the Earth’s global geochemical water cycle is the amount of water that enters the deep mantle through the subduction and recycling of hydrated oceanic lithosphere. Here we address the question of recycling of water into the deep mantle by characterizing the volatile contents of different mantle components as sampled by ocean island basalts and mid-oceanridge basalts. Although all mantle plume (ocean island) basalts seem to contain more water than mid-ocean-ridge basalts, we demonstrate that basalts associated with mantle plume components containing subducted lithosphere—‘enriched-mantle’ or ‘EMtype’ basalts—contain less water than those associated with a common mantle source. We interpret this depletion as indicating that water is extracted from the lithosphere during the subduction process, with greater than 92 per cent efficiency. Quantifying the Earth’s global geochemical water cycle is important for modelling physical properties of mantle materials1, melt generation2 and planetary evolution3. Water is added to oceanic crust at or near mid-ocean-ridge crests via hydrothermal alteration, and some or most of this water is lost during dehydration of subducting lithosphere, contributing to flux-melting of the overlying mantle wedge, generation of island arc magmas, and addition of water to the atmosphere–ocean system4. However, the amount of water that survives subduction and is transported to the deep mantle, perhaps contributing to subsequent ‘wet’ plume generation and ocean island volcanism5, is poorly known.
Recycling of crustal materials and mantle heterogeneity Chemical diversity within the mantle is closely related to recycling of oceanic crust and sediments into the mantle by the plate tectonic process of subduction6. On the basis of radiogenic Pb, Nd and Sr isotopes in ocean island basalts, five isotopically extreme mantle components have been defined7–9: (1) DMM (depleted MORB mantle); (2) HIMU (‘high m’ where m ¼ 238U/204Pb); (3) EM1 (enriched mantle 1); (4) EM2 (enriched mantle 2); and (5) LOMU (‘low m’) with very low time-integrated U/Pb ratio; MORB indicates mid-ocean-ridge basalt. In this context, ‘enriched’ refers to time-integrated Rb/Sr, Sm/Nd and/or (U þ Th)/Pb ratios higher than primitive mantle (bulk silicate Earth). The origin and location of these end-members are actively debated, but it is generally accepted that the shallow upper mantle is the source of depleted MORB. HIMU is believed to represent recycled, old, hydrothermally altered oceanic crust, dehydrated and metamorphosed to eclogite during subduction. HIMU sources are depleted in Pb and K and enriched in U, Nb and Ta (refs 10, 11) relative to MORB. EM1 includes either recycled oceanic crust plus a few per cent pelagic sediment10–12 or metasomatized subcontinental lithosphere13. EM2 includes recycled oceanic crust containing a few per cent of continent-derived sediment11. LOMU is distinguished from EM1 by its unusually high 87Sr/86Sr and 207Pb/204Pb ratios, and very low 143Nd/144Nd and 177Hf/176Hf ratios, suggesting an ancient continental source9,14,15. Here we will refer to EM1, EM2, LOMU and mixtures of these simply as EM components, as all seem to include recycled materials of continental derivation. Whereas the five end-members constrain the compositional NATURE | VOL 420 | 28 NOVEMBER 2002 | www.nature.com/nature
extremes, most data from ocean islands form linear mixing arrays that converge in a limited region of isotopic space that has been given various acronyms, including C (‘common’ mantle component; ref. 16) and FOZO (‘focus zone’; refs 17, 18), used here. FOZO has moderately depleted Sr and Nd signatures, radiogenic Pb isotopes, and elevated 3He/4He ratios18. The high He isotope ratios of FOZO suggest an origin in the lower mantle16,19. However, radiogenic isotopic compositions and trace-element ratios are inconsistent with FOZO being primitive mantle, and suggest involvement of a HIMU component6,10,20–25. It has also been proposed that EM components are associated with mantle metasomatism26, implying concentrations of water and/or carbon dioxide greater than those in ‘normal’ mantle, if the metasomatic agent is a hydrous or carbonate phase. However, Dixon and Clague27 showed that H2O/Ce in Loihi summit glasses correlates negatively with 206Pb/204Pb, and interpreted this as evidence for a relatively ‘dry’ EM component in the Hawaiian plume. Here we report H2O and CO2 concentration data (measured by transmission infrared spectroscopy on MORB glasses from ridges influenced by various mantle plume components in the North and South Atlantic oceans); our aim is to see if the Hawaii result is unique, or if it represents a global EM characteristic.
H2O/Ce as a tracer of magmatic and nonmagmatic fractionations Previous workers28,29 have found it useful to compare water concentrations to concentrations of elements or oxides of similar incompatibility, such as K2O, P2O5, La and Ce. Water has a bulk distribution coefficient (D H2O) of ,0.01 (refs 28, 29), slightly higher than K2O and La and lower than Ce and P2O5. For virtually all chemical species in oceanic basalts, ratios of more-incompatible to less-incompatible elements or oxides correlate positively. Thus H2O ratioed to a slightly more compatible species (Ce, P2O5) correlates positively with other trace-element ratios indicative of source region enrichment or depletion (such as (La/Sm)n, K2O/ TiO2), as has been observed at the Easter microplate (EMP) and at the Easter–Salas y Gomez seamount chain (ESC)30, and for the East Pacific Rise and near-rise seamounts29. These correlations are thought to represent mantle source characteristics, because as long as the extent of melting (F) is greater than the bulk distribution coefficient (D), trace-element ratios of incompatible elements in
© 2002 Nature Publishing Group
385
articles basaltic melts will be similar to those of the source region. These data show that for areas such as the Easter microplate, water behaves as other incompatible elements and appears to be controlled by magmatic processes. Shallow level processes introduce scatter into these correlations through degassing (which decreases H2O/Ce) or assimilation of seawater-derived components (which increases H2O/Ce). These modifications of water concentrations can be identified by examining
water concentration as a function of partial melting, differentiation, eruption depth, vesiculation, and chlorine concentration27–32. For example, to correct for water degassing, bulk water concentrations can be estimated by adding the amount of water in vesicles to that dissolved in the glass27,31. Once shallow level processes have been accounted for, deeper processes can be investigated. For example, addition of hydrous fluid phases by mantle metasomatism, or addition of a mantle component containing hydrous phases, should increase H2O/Ce, whereas incorporation of severely dehydrated components will decrease H2O/Ce. If variations in H2O/Ce correlate with radiogenic isotopic compositions, then the hydration or dehydration event probably occurred at the same time as the fractionation between parent and daughter elements.
Water in plume-influenced MORB from the South Atlantic
Figure 1 Comparison of radiogenic isotopic compositions, H2O, K2O and H2O/Ce in basaltic glasses. Data are from the Discovery and Shona anomalies9 along the southern Mid-Atlantic Ridge, and from the Easter microplate and Easter–Salas y Gomez seamount chain (EMP-ESC) in the southeast Pacific35,36. a, 87Sr/86Sr versus 206Pb/204Pb. Isotopic compositions for mantle end-members are from ref. 51. Open symbols (Discovery and Shona group 2) represent samples having an EM-type mantle plume component. Filled symbols (Shona group 1 and EMP-ESC) represent samples containing a FOZO-type or mildly-HIMU plume component. b, H2O versus K2O. Water concentrations increase with increasing K2O concentrations; however, EM-type basalts contain less water at a given K2O than C-type basalts. c, H2O/Ce versus 87Sr/86Sr for the South Atlantic and Easter microplate. EM-type basalts have significantly lower H2O/Ce than FOZO-type basalts. Several samples have high H2O/Ce ratios (.300) that do not correlate with major- or trace-element or isotopic compositions. These samples may have assimilated water during transport and eruption. Symbols are as in b. 386
South Atlantic samples were dredged along the Mid-Atlantic Ridge between 408 S and 548 S during the RV Maurice Ewing EW93-09 cruise (see Supplementary Table S1). This region contains two distinct geochemical, gravity and bathymetric anomalies. The northern ‘Discovery’ anomaly is centred at 47.58 S and is influenced by the Discovery plume9,33. The southern ‘Shona’ anomaly is centred at 51.58 S (refs 9, 33, 34). We compare the South Atlantic samples to EMP and ESC glasses because both areas involve plume–ridge interaction, and Pacific and South Atlantic MORB have similar regional H2O/Ce ratios23, and thus similar depleted end-member compositions. Radiogenic isotopic data for EMP-ESC basalts (Fig. 1a) have been interpreted as the result of binary mixing between depleted MORB and mildlyHIMU FOZO (Salas y Gomez plume) sources35,36. Isotopic compositions of Discovery glasses show a wide range of 87 Sr/86Sr at a relatively constant 206Pb/204Pb, consistent with pseudo-binary mixing of DMM, FOZO and an EM component, specifically LOMU9 (Fig. 1a). Contribution of this EM component is least for samples 26, 28, 34 and 40, and greatest for sample 7, 9, 33 and 25 (ref. 37)37. Whether this component is recycled in the upper or lower mantle is not known. However, the nearly solar 21Ne/22Ne ratios and high 3He/4He in some of the Discovery MORB glasses (samples 2–5) suggest recycling at least to the 660-km discontinuity, where more primitive mantle could have been entrained by the Discovery plume38. Shona glasses form two groups in Sr–Nd–Pb–Hf isotopic space. Shona group 1 samples (dredges 11, 14, 15, 17 and 19) show mixing between DMM and FOZO with mild HIMU influence, similar to the Easter trend. Shona group 2 samples (dredges 18, 20, 21, 22 and 23) show additional influence of an EM component (EM1-EM2) thought to represent recycled oceanic crust þ pelagic sediments15. Water correlates positively with K2O for the South Atlantic, but with very different slopes defined by Shona and Discovery samples (Fig. 1b). Linear correlations between H2O and K2O for Shona group 1, Easter, and other South Atlantic (31-468S) MORB samples are indistinguishable (Fig. 1b). One outlier (sample 19) has more water at a given K2O (about 0.1 wt% extra) than samples forming the main trend, with H2O/Ce similar to several ESC samples, possibly related to serpentine assimilation during ascent23. For the remaining samples, water is more compatible than K and less compatible than Ce, consistent with the behaviour of these elements in magmatic systems. Thus, the characteristics of the plume and DMM end-members for the Shona group 1 samples are similar to those of the EMP-ESC30, for which the FOZO component has H2O/ K2O ¼ ,0.8 and H2O/Ce ¼ 210 ^ 30 and the DMM component has H2O/K2O ¼ ,5 and H2O/Ce ¼ 150 ^ 10 (Fig. 1c). Shona group 2 and Discovery samples (87Sr/86Sr . 0.703) that contain an EM component have distinctly lower H2O concentrations and lower H2O/Ce than Shona group 1 and Easter samples. Shona group 2 data form trends sub-parallel to Shona group 1 data but at progressively lower H2O content and 206Pb/204Pb. The
© 2002 Nature Publishing Group
NATURE | VOL 420 | 28 NOVEMBER 2002 | www.nature.com/nature
articles Discovery samples form a distinct binary mixing line converging with that of the Easter group at the depleted source end (Fig. 1b, c). Discovery samples with the most definitive EM signature have H2O/ K2O half that of the Easter samples. The observation that sample 25 has relatively low water concentration and H2O/Ce is very important. Even though it has the highest 87Sr/86Sr ratio, reflecting a mantle source with time-integrated enrichment of incompatible elements (for example, high Rb/Sr), it is not the most enriched in incompatible elements, perhaps related to a recent episode of
melting and melt extraction39. Sample 25 has only 0.15 wt% water. In order for this sample to have degassed significant quantities of water, it would have to have been erupted at or below the water solubility pressure of ,3 bar (30 m water depth), much shallower than its actual eruption depth at 2,032 m. It is almost certain that its low water concentration is a primary magmatic characteristic. Samples most enriched in 87Sr/86Sr (7, 25 and 33) have the lowest H2O/ K2O. If these ratios were caused by the same type of (magmatic) process as observed for the Easter samples, then these samples should also have the highest H2O/Ce. Instead, samples with the lowest H2O/K2O have low H2O/Ce, suggesting control by a different process that selectively lowers water content. Thus, the EM components in the mantle sources for the Shona and Discovery samples are associated with lower water concentrations and lower ratios of water to similarly incompatible elements compared to the common plume component (FOZO).
Water in plume-influenced MORB from the North Atlantic
Figure 2 Comparison of radiogenic isotopic compositions, H2O, K2O and H2O/Ce in basaltic glasses from the North Atlantic. a, 87Sr/86Sr versus 206Pb/204Pb. Azores platform basalts43 (black filled triangles) are consistent with mixing between DMM and C-type mantle components. Great Meteor anomaly basalts (blue open circles) are consistent mixing between EM-type and FOZO-type mantle components. b, H2O versus K2O. K2O data from ref. 40. Measured dissolved H2O concentrations are shown as black half-filled squares (Azores platform) and blue crossed squares (Great Meteor). Calculated bulk H2O concentrations (see Supplementary Tables S1 and S2) are shown as black filled triangles (Azores platform) and blue open circles (Great Meteor). Lines connect dissolved and bulk water concentrations for individual samples. Calculated bulk H2O concentrations for dredge 17 from the summit of the Azores platform are probably low owing to open-system degassing (J.E.D., manuscript in preparation). EM-type basalts have less water at a given K2O content than FOZO-type basalts from the North Atlantic. c, H2O/Ce versus 87Sr/86Sr for the North Atlantic. EM-type basalts have lower H2O/Ce than FOZO-type basalts and depleted MORB. Symbols as in b. NATURE | VOL 420 | 28 NOVEMBER 2002 | www.nature.com/nature
The north Mid-Atlantic Ridge samples were dredged by the RV Atlantis II (FAZAR expedition40) between 33.28 and 40.58 N with depths varying between 926 and 3,900 m (see Supplementary Table S2). Two distinct geochemical anomalies occur within this region, including the long-wavelength anomaly associated with the intersection of the Azores archipelago with the Mid-Atlantic Ridge at ,38–398 N (Azores platform) and a short-wavelength anomaly centred at 358 N (the Great Meteor anomaly41, used here, or the ‘north of Oceanographer fracture zone anomaly’42). Sr and Pb isotopic data43 (Fig. 2a) show that Azores platform basalts form a linear mixing array between DMM and FOZO. In contrast, basalts from the Great Meteor anomaly form a mixing array between FOZO and an EM component, with samples 9-10, 45-3 and 44-1 having the highest 87Sr/86Sr ratios. An exception is 49-3, which falls within the Azores platform array. Because of the high vesicularity of many of the Azores platform and Great Meteor samples, loss of water into vesicles may have been significant. Both measured dissolved and calculated bulk water concentrations (see Supplementary Information) are shown in Fig. 2b and c. Samples from dredges 17, 19, 21 and 22 from the summit of the Azores platform (926–1,950 m) are highly vesicular (25–60 vol.%) and have lost about 10–30% of their initial water into vesicles. Dissolved water concentrations in the less vesicular Azores platform samples correlate positively with K2O and have H2O/Ce ratios (253 ^ 33) consistent with previously reported values for the North Atlantic28. Estimated bulk H2O/Ce for the highly vesicular samples (247 ^ 54) are similar to values for nonvesicular samples for all but the shallowest dredge (17), which may have suffered some gas loss (open-system degassing). Great Meteor anomaly samples are also highly vesicular (5– 60 vol.%), with vesicularity correlating roughly inversely with depth (5 vol.% at 3,900 m to 60 vol.% at 1,657 m). Samples with the largest EM component have dissolved and bulk water concentrations at a given K2O (Fig. 2b) that are significantly lower than undegassed samples from the Azores platform. H2O/Ce is also 30% lower than the mean for the Azores platform samples (168 ^ 29 versus 253 ^ 33). Though the correlation is rough, extrapolation from FOZO through the Great Meteor data yields H2O/Ce , 100 at 87 Sr/86Sr . 0.705. Thus, like the Discovery samples, EM components in the North Atlantic are depleted in water relative to FOZO.
Discussion In each setting described above, there are two distinct mantle plume components: one close to FOZO (common mantle plume component), and the other enriched in recycled material of continental derivation (EM). In each case, the ratio of water to other similarly incompatible elements is lower in the EM component than in FOZO. Our data show that basalts influenced by EM mantle
© 2002 Nature Publishing Group
387
articles components are depleted in water (‘damp’ plumes) relative to those influenced by FOZO (‘wet’ plumes). What is FOZO?
We have shown that FOZO plumes are ‘wetter’ than EM plumes, but we still need to address the origin of H2O in these mantle components. Though the origin of FOZO is an area of active debate, the following characteristics are generally accepted. First, FOZO is not pristine primitive mantle. Paradoxically, Sr- and Nd-isotopic compositions require moderate time-integrated depletion in incompatible elements (Rb/Sr lower and Sm/Nd higher than primitive mantle), whereas the present-day source is enriched in incompatible elements6,44. In addition, Pb-, Os- and O-isotopic compositions and trace-element ratios (Nb/U, Nb/Th and Ce/Pb) suggest involvement of a HIMU component6,10,21–25,44. Second, high concentrations of Ni, Cr and heavy rare-earth elements in FOZO require melting of a peridotitic source, and are inconsistent with melting of a purely eclogitic (HIMU) source44,45. Third, FOZO typically has elevated 3He/4He, indicating involvement of mantle that is less degassed and more primitive than MORB mantle, probably located deeper than the MORB source17,18 . High 3 He/4He also rules out generation of FOZO by mixing between DMM and HIMU components45. Lastly, the globally widespread occurrence of FOZO in ocean island basalts and carbonatites makes it unlikely that FOZO is the product of melting heterogeneous recycled crustal materials that make up only 4–10% of mantle volume18,45. Similar to a recent model by Donnelly et al.44, we suggest that FOZO represents dominantly primitive or slightly depleted peridotitic mantle (,97–99%), metasomatized by a small amount (1–3%) of a small extent melt (2–6%) of a HIMU component. If the lower mantle is composed of peridotite with ‘streaks’ of recycled eclogitized oceanic crust (HIMU), then this metasomatism could occur during ascent of plumes, a process recently proposed to explain major-element heterogeneity in picrites and komatiites46. In such a mixture, concentrations of noble gases and compatible elements (Ni, Cr and heavy rare-earth elements) in the FOZO source come almost completely from the peridotitic mantle, whereas roughly half to two-thirds of the highly incompatible elements are supplied by the metasomatic HIMUmelt. Thus, water in FOZO plumes is a combination of primitive and recycled water. H2O in mantle sources of FOZO and HIMU plumes
On the basis of eclogite measurements47, subducted oceanic crust
Table 1 Dehydration efficiency Rock
H2O/Ce
H2O (wt%)
250 150 2,500–5,000 1,280
0.3 0.2 2–3 7.3
Ce (p.p.m.)‡
Dehydration (%)
.............................................................................................................................................................................
Crustal components: Pre-subduction Atlantic MORB* 12 Pacific MORB* 14 Mature oceanic crust* ,6 Global subducted sediment† 57 Crustal components: Post-subduction Mature oceanic crust ,6 Global subducted sediment 57 Mantle components DMM* 0.5 Atlantic FOZOk 3 Pacific FOZO{ 3.8 EM# 4
,100 ,100
0.06§ ,0.57
200 250 200 ,100
0.010 0.075 0.075 ,0.04
97 .92
............................................................................................................................................................................. * GERM (Geochemical Earth Reference Model) database (http://earthref.org/databases/GERMRD/ main.htm). † GLOSS (Global Subducting Sediment)48. ‡ Represents conservative estimate—some rare-earth elements are lost from subducting lithosphere during subduction. § Bulk water concentration in eclogite47. kAsimow, P., C.L. and J.E.D., manuscript in preparation. {From refs 27 and 30. # EM sources have Ce contents similar to HIMU sources6 and slightly greater than FOZO-type sources. EM mantle sources have roughly half as much water as FOZO-type sources (this study).
388
(HIMU source) is low in water, with ,600 p.p.m. H2O, ,6 p.p.m. Ce, and H2O/Ce ¼ ,100. These data suggest that subducted oceanic crust is efficiently (,97%) dehydrated during subduction (Table 1). In contrast, mantle sources of FOZO plumes have ,750 p.p.m. H2O and H2O/Ce of ,210 in the Pacific and South Atlantic and ,250 in the North Atlantic (ref. 25, and P. Asimow, C.L. and J.E.D., manuscript in preparation). The high H2O/Ce in FOZO plumes cannot be derived from recycled oceanic crust; a significant amount of water must be juvenile, left over from planetary accretion. We believe that it is the presence of this juvenile water that makes FOZO plumes ‘wet’. Water in EM components
We have shown that recycled oceanic crust is efficiently dehydrated during the subduction process. The lower H2O/Ce of EM plumes relative to FOZO plumes indicates that recycled continentally derived components are also efficiently dehydrated. For example, an estimate of globally subducted sediment48 contains 57 p.p.m. Ce and ,7.3 wt% H2O (H2O/Ce ¼ 1,280). Reduction of the H2O/Ce to ,100, lower than the ambient upper mantle of .150, requires .92% dehydration of the water from sediments (Table 1). These dehydrated sediments are still fairly hydrous (,0.57 wt% H2O), but addition of only a few per cent sediments to HIMU will only increase the water concentration by 50 to 150 p.p.m. Thus our data suggest that dehydration is efficient for all recycled materials. These estimates of dehydration are consistent with other approaches that have considered mass balance at the subduction zone itself49. Discovery of dense hydrous magnesian silicate phases in laboratory experiments is generally considered consistent with a scheme whereby water may be efficiently transported into the deep mantle during subduction50. However, laboratory generation of phases capable of holding significant amounts of water does not constitute evidence that these phases in fact carry significant water to the deep mantle. Our data suggest that much of the recycled material that has passed through the subduction ‘factory’ has H2O/Ce , 100. Thus, as is the case for Pb (refs 21–23), the subduction process effectively partitions water into the exosphere (mantle wedge, crust, ocean and atmosphere), resulting in time-integrated depletion of water relative to other incompatible elements in the recycled lithosphere and accompanying continental-derived materials. Thus, like a wellsqueezed sponge, the residual slab and EM plume components are not ‘wet’, but ‘damp’. A Received 28 August; accepted 9 October 2002; doi:10.1038/nature01215. 1. Hirth, G. & Kohlstedt, D. L. Water in the oceanic upper mantle: implications for rheology, melt extraction and the evolution of the lithosphere. Earth Planet. Sci. Lett. 144, 93–108 (1996). 2. Hirose, K. & Kawamoto, T. Hydrous partial melting of lherzolite at 1 GPa; the effect of H2O on the genesis of basaltic magmas. Earth Planet. Sci. Lett. 133, 463–473 (1995). 3. Abe, Y., Ohtani, E., Okuchi, T., Righter, K. & Drake, M. Origin of the Earth and Moon (eds Canup, R. M. & Righter, K.) 413–433 (Univ. Arizona Press, Tucson, 2000). 4. Schmidt, M. W. & Poli, S. Experimental base water budgets for dehydrating slabs and consequences for arc magma generation. Earth Planet. Sci. Lett. 163, 361–379 (1998). 5. Schilling, J.-G., Bergeron, M. B. & Evans, R. Halogens in the mantle beneath the North Atlantic. Phil. Trans. R. Soc. Lond. A 297, 147–178 (1980). 6. Hofmann, A. W. Mantle geochemistry: the message from oceanic volcanism. Nature 385, 219–229 (1997). 7. White, W. M. Sources of oceanic basalts: Radiogenic isotopic evidence. Geology 13, 115–118 (1985). 8. Zindler, A. & Hart, S. Chemical geodynamics. Annu. Rev. Earth Planet. Sci. 14, 493–571 (1986). 9. Douglass, J., Schilling, J.-G. & Fontignie, D. Plume-ridge interactions of the Discovery and Shona mantle plumes with the southern Mid-Atlantic Ridge (408-558S). J. Geophys. Res. 104, 2941–2962 (1999). 10. Weaver, B. L. The origin of ocean island basalt end-member compositions: Trace element and isotopic constraints. Earth Planet. Sci. Lett. 204, 381–397 (1991). 11. Chauvel, C., Hofmann, A. W. & Vidal, P. HIMU-EM: the French Polynesian connection. Earth Planet. Sci. Lett. 110, 99–119 (1992). 12. Blichert-Toft, J., Frey, F. A. & Albare`de, F. Hf isotope evidence for pelagic sediments in the source of Hawaiian basalts. Science 285, 879–882 (1999). 13. McKenzie, D. & O’Nions, R. K. Mantle reservoirs and ocean island basalts. Nature 301, 229–231 (1983). 14. Kamenetsky, V. S. et al. Remnants of Gondwanan continental lithosphere in oceanic upper mantle: Evidence from the South Atlantic Ridge. Geology 29, 243–246 (2001). 15. Andres, M., Blichert-Toft, J. & Schilling, J.-G. Hafnium isotopes in basalts from the Southern MidAtlantic Ridge from 408S to 558S: Trace element evidence for contamination of ambient asthenospheric mantle. Earth Planet. Sci. Lett. (in the press).
© 2002 Nature Publishing Group
NATURE | VOL 420 | 28 NOVEMBER 2002 | www.nature.com/nature
articles 16. Hanan, B. B. & Graham, D. W. Lead and helium isotope evidence from oceanic basalts for a common deep source of mantle plumes. Science 272, 991–995 (1996). 17. Hart, S. R., Hauri, E. H., Oschmann, L. A. & Whitehead, J. A. Mantle plumes and entrainment: The isotopic evidence. Science 256, 517–520 (1992). 18. Hauri, E. H., Whitehead, J. A. & Hart, S. R. Fluid dynamic and geochemical aspects of entrainment in mantle plumes. J. Geophys. Res. 99, 24275–24300 (1994). 19. Alle`gre, C. J., Hofmann, A. W. & O’Nions, R. K. The argon constraints on mantle structure. Geophys. Res. Lett. 23, 3555–3557 (1996). 20. Hofmann, A. W. Nb in Hawaiian magma: constraints on source composition and evolution. Chem. Geol. 57, 17–30 (1986). 21. Hofmann, A. W., Jochum, K. P., Seufert, M. & White, W. M. Nb and Pb in oceanic basalts: new constraints on mantle evolution. Earth Planet. Sci. Lett. 79, 33–45 (1986). 22. Miller, D. M., Goldstein, S. L. & Langmuir, C. H. Cerium/lead and lead isotope ratios in arc magmas and the enrichment of lead in the continents. Nature 368, 514–520 (1994). 23. Chauvel, C., Goldstein, S. L. & Hofmann, A. W. Hydration and dehydration of oceanic crust controls Pb evolution in the mantle. Chem. Geol. 126, 65–75 (1995). 24. Shirey, S. B. & Walker, R. J. The Re-Os isotope system in cosmochemistry and high-temperature geochemistry. Annu. Rev. Earth Planet. Sci. 26, 423–500 (1998). 25. Eiler, J. M. Stable Isotopes in Igneous Processes (eds Cole, D. and Valley, J. W.) (Mineralological Society of America, Washington DC, 2002). 26. Menzies, M. & Murthy, V. R. Enriched mantle: Nd and Sr isotopes in diopsides from kimberlite nodules. Nature 283, 634–636 (1980). 27. Dixon, J. E. & Clague, D. A. Volatiles in basaltic glasses from Loihi Seamount, Hawaii: Evidence for a relatively dry plume component. J. Petrol. 42, 627–654 (2001). 28. Michael, P. J. Regionally distinctive sources of depleted MORB: Evidence from trace elements and H2O. Earth Planet. Sci. Lett. 131, 301–320 (1995). 29. Danyuschevsky, L. V., Eggins, S. M., Falloon, T. J. & Christie, D. M. H2O abundance in depleted to moderately enriched mid-ocean ridge magmas; Part I: Incompatible behaviour, implications for mantle storage, and origin of regional variations. J. Petrol. 41, 1329–1364 (2000). 30. Simons, K., Dixon, J. E., Schilling, J.-G., Kingsley, R. & Poreda, R. Volatiles in basaltic glasses from the Easter-Salas y Gomez Seamount Chain and Easter Microplate: Implications for geochemical cycling of volatile elements. Geochem. Geophys. Geosyst. 3, doi: 10.1029/2001GC000173 (2002). 31. Dixon, J. E., Clague, D. A., Wallace, P. & Poreda, R. Volatiles in alkalic basalts from the North Arch Volcanic Field, Hawaii: Extensive degassing of deep submarine-erupted alkalic series lavas. J. Petrol. 38, 911–939 (1997). 32. Michael, P. J. & Cornell, W. C. Influence of spreading rate and magma supply on crystallization and assimilation beneath mid-ocean ridges: Evidence from chlorine and major element chemistry of midocean ridge basalts. J. Geophys. Res. 103, 18325–18356 (1998). 33. Douglass, J., Schilling, J.-G. & Kingsley, R. H. Influence of the Discovery and Shona mantle plumes on the southern Mid-Atlantic Ridge: Rare earth evidence. Geophys. Res. Lett. 22, 2893–2896 (1995). 34. le Roex, A. P., Dick, H. J. B., Gulen, L., Reid, A. M. & Erlank, A. J. Local and regional heterogeneity in MORB from the Mid-Atlantic Ridge between 54.58S and 518S: Evidence for geochemical enrichment. Geochim. Cosmochim. Acta 51, 541–555 (1987). 35. Kingsley, R. H. & Schilling, J.-G. Plume-ridge interaction in the Easter-Salas y Gomez seamount chain-Easter Microplate system: Pb isotope evidence. J. Geophys. Res. 103, 24159–24177 (1998). 36. Cheng, Q. C., Macdougall, J. D. & Zhu, P. Isotopic constraints on the Easter Seamount Chain source. Contrib. Mineral. Petrol. 135, 225–233 (1999).
NATURE | VOL 420 | 28 NOVEMBER 2002 | www.nature.com/nature
37. Douglass, J. & Schilling, J.-G. Systematics of three-component, pseudo-binary mixing lines in 2D isotope ratio space representations and implications for mantle plume-ridge interaction. Chem. Geol. 163, 1–23 (2000). 38. Sarda, P., Moreira, M., Staudacher, T., Schilling, J.-G. & Alle`gre, C. J. Rare gas systematics on the southernmost Mid-Atlantic Ridge: Constraints on the lower mantle and the Dupal source. J. Geophys. Res. 105, 5973–5996 (2000). 39. le Roux, P. J. et al. Mantle heterogeneity beneath the southern Mid-Atlantic Ridge: trace element evidence for contamination of ambient asthenospheric mantle. Earth Planet. Sci. Lett. (in the press). 40. Langmuir, C. H., Klinkhammer, G. & Bougalult, H. Cruise Report: Evaluation of the Relationships among Segmentation, Hydrothermal Activity and Petrological Diversity on the Mid-Atlantic Ridge (FAZAR) Technical Report LDEO-92-3 (Lamont-Doherty Earth Observatory, Columbia Univ., Palisades, 1992). 41. Rideout, M. L. & Schilling, J.-G. Rare-earth elements, 87Sr/86Sr, and 143Nd/144Nd mantle source variations. Init. Rep. DSDP 82, 483–496 (1985). 42. Shirey, S. B., Bender, J. F. & Langmuir, C. H. Three-component isotopic heterogeneity near the Oceanographer transform, Mid-Atlantic Ridge. Nature 325, 217–223 (1987). 43. Dosso, L. et al. The age and distribution of mantle heterogeneity along the Mid-Atlantic Ridge (31418N). Earth Planet. Sci. Lett. 170, 269–286 (1999). 44. Donnelly, K., Langmuir, C. H., Goldstein, S. L. & Lagatta, A. The origin of alkali and ocean island basalts: contradictions and solutions. Eos 82, F1402 (2001). 45. Bell, K. & Tilton, G. R. Probing the mantle: The story from carbonatites. Eos 83, 273–277 (2002). 46. Gibson, S. A. Major element heterogeneity in Archean to Recent mantle plume starting-heads. Earth Planet. Sci. Lett. 195, 59–74 (2002). 47. Nadeau, S., Philippot, P. & Pineau, F. Fluid inclusion and mineral isotopic compositions (H-C-O) in eclogitic rocks as tracers of local fluid migration during high-pressure metamorphism. Earth Planet. Sci. Lett. 114, 431–448 (1993). 48. Plank, T. & Langmuir, C. H. The chemical composition of subducting sediment and its consequences for the crust and mantle. Chem. Geol. 145, 325–394 (1998). 49. Peacock, S. M. Fluid processes in subduction zones. Science 248, 329–337 (1990). 50. Thompson, A. B. Water in the Earth’s upper mantle. Nature 358, 295–302 (1992). 51. Schilling, J.-G. Les Isotopes Radioge´niques en Ge´ologie: Chronologie Geochimie (ed. Duthou, J. L.) 1–34 (Socie´te´ Franc¸aise de Mine´ralogie et de Cristallographie, Paris, 1992).
Supplementary Information accompanies the paper on Nature’s website (ç http://www.nature.com/nature). Acknowledgements This research was funded by the NSF. A review by N. Arndt improved the manuscript. Competing interests statement The authors declare that they have no competing financial interests. Correspondence and requests for materials should be addressed to J.E.D. (e-mail: [email protected]).
© 2002 Nature Publishing Group
389
Recycled dehydrated lithosphere observed in plume-influenced mid-ocean-ridge basalt Jacqueline Eaby Dixon*, Loretta Leist*, Charles Langmuir† & Jean-Guy Schilling‡ * RSMAS/MGG, University of Miami, 4600 Rickenbacker Causeway, Miami, Florida 33149, USA † Department of Earth & Planetary Sciences, Harvard University, Cambridge, Massachusetts 02138, USA ‡ Graduate School of Oceanography, University of Rhode Island, Narragansett, Rhode Island 02882, USA
...........................................................................................................................................................................................................................
A substantial uncertainty in the Earth’s global geochemical water cycle is the amount of water that enters the deep mantle through the subduction and recycling of hydrated oceanic lithosphere. Here we address the question of recycling of water into the deep mantle by characterizing the volatile contents of different mantle components as sampled by ocean island basalts and mid-oceanridge basalts. Although all mantle plume (ocean island) basalts seem to contain more water than mid-ocean-ridge basalts, we demonstrate that basalts associated with mantle plume components containing subducted lithosphere—‘enriched-mantle’ or ‘EMtype’ basalts—contain less water than those associated with a common mantle source. We interpret this depletion as indicating that water is extracted from the lithosphere during the subduction process, with greater than 92 per cent efficiency. Quantifying the Earth’s global geochemical water cycle is important for modelling physical properties of mantle materials1, melt generation2 and planetary evolution3. Water is added to oceanic crust at or near mid-ocean-ridge crests via hydrothermal alteration, and some or most of this water is lost during dehydration of subducting lithosphere, contributing to flux-melting of the overlying mantle wedge, generation of island arc magmas, and addition of water to the atmosphere–ocean system4. However, the amount of water that survives subduction and is transported to the deep mantle, perhaps contributing to subsequent ‘wet’ plume generation and ocean island volcanism5, is poorly known.
Recycling of crustal materials and mantle heterogeneity Chemical diversity within the mantle is closely related to recycling of oceanic crust and sediments into the mantle by the plate tectonic process of subduction6. On the basis of radiogenic Pb, Nd and Sr isotopes in ocean island basalts, five isotopically extreme mantle components have been defined7–9: (1) DMM (depleted MORB mantle); (2) HIMU (‘high m’ where m ¼ 238U/204Pb); (3) EM1 (enriched mantle 1); (4) EM2 (enriched mantle 2); and (5) LOMU (‘low m’) with very low time-integrated U/Pb ratio; MORB indicates mid-ocean-ridge basalt. In this context, ‘enriched’ refers to time-integrated Rb/Sr, Sm/Nd and/or (U þ Th)/Pb ratios higher than primitive mantle (bulk silicate Earth). The origin and location of these end-members are actively debated, but it is generally accepted that the shallow upper mantle is the source of depleted MORB. HIMU is believed to represent recycled, old, hydrothermally altered oceanic crust, dehydrated and metamorphosed to eclogite during subduction. HIMU sources are depleted in Pb and K and enriched in U, Nb and Ta (refs 10, 11) relative to MORB. EM1 includes either recycled oceanic crust plus a few per cent pelagic sediment10–12 or metasomatized subcontinental lithosphere13. EM2 includes recycled oceanic crust containing a few per cent of continent-derived sediment11. LOMU is distinguished from EM1 by its unusually high 87Sr/86Sr and 207Pb/204Pb ratios, and very low 143Nd/144Nd and 177Hf/176Hf ratios, suggesting an ancient continental source9,14,15. Here we will refer to EM1, EM2, LOMU and mixtures of these simply as EM components, as all seem to include recycled materials of continental derivation. Whereas the five end-members constrain the compositional NATURE | VOL 420 | 28 NOVEMBER 2002 | www.nature.com/nature
extremes, most data from ocean islands form linear mixing arrays that converge in a limited region of isotopic space that has been given various acronyms, including C (‘common’ mantle component; ref. 16) and FOZO (‘focus zone’; refs 17, 18), used here. FOZO has moderately depleted Sr and Nd signatures, radiogenic Pb isotopes, and elevated 3He/4He ratios18. The high He isotope ratios of FOZO suggest an origin in the lower mantle16,19. However, radiogenic isotopic compositions and trace-element ratios are inconsistent with FOZO being primitive mantle, and suggest involvement of a HIMU component6,10,20–25. It has also been proposed that EM components are associated with mantle metasomatism26, implying concentrations of water and/or carbon dioxide greater than those in ‘normal’ mantle, if the metasomatic agent is a hydrous or carbonate phase. However, Dixon and Clague27 showed that H2O/Ce in Loihi summit glasses correlates negatively with 206Pb/204Pb, and interpreted this as evidence for a relatively ‘dry’ EM component in the Hawaiian plume. Here we report H2O and CO2 concentration data (measured by transmission infrared spectroscopy on MORB glasses from ridges influenced by various mantle plume components in the North and South Atlantic oceans); our aim is to see if the Hawaii result is unique, or if it represents a global EM characteristic.
H2O/Ce as a tracer of magmatic and nonmagmatic fractionations Previous workers28,29 have found it useful to compare water concentrations to concentrations of elements or oxides of similar incompatibility, such as K2O, P2O5, La and Ce. Water has a bulk distribution coefficient (D H2O) of ,0.01 (refs 28, 29), slightly higher than K2O and La and lower than Ce and P2O5. For virtually all chemical species in oceanic basalts, ratios of more-incompatible to less-incompatible elements or oxides correlate positively. Thus H2O ratioed to a slightly more compatible species (Ce, P2O5) correlates positively with other trace-element ratios indicative of source region enrichment or depletion (such as (La/Sm)n, K2O/ TiO2), as has been observed at the Easter microplate (EMP) and at the Easter–Salas y Gomez seamount chain (ESC)30, and for the East Pacific Rise and near-rise seamounts29. These correlations are thought to represent mantle source characteristics, because as long as the extent of melting (F) is greater than the bulk distribution coefficient (D), trace-element ratios of incompatible elements in
© 2002 Nature Publishing Group
385
articles basaltic melts will be similar to those of the source region. These data show that for areas such as the Easter microplate, water behaves as other incompatible elements and appears to be controlled by magmatic processes. Shallow level processes introduce scatter into these correlations through degassing (which decreases H2O/Ce) or assimilation of seawater-derived components (which increases H2O/Ce). These modifications of water concentrations can be identified by examining
water concentration as a function of partial melting, differentiation, eruption depth, vesiculation, and chlorine concentration27–32. For example, to correct for water degassing, bulk water concentrations can be estimated by adding the amount of water in vesicles to that dissolved in the glass27,31. Once shallow level processes have been accounted for, deeper processes can be investigated. For example, addition of hydrous fluid phases by mantle metasomatism, or addition of a mantle component containing hydrous phases, should increase H2O/Ce, whereas incorporation of severely dehydrated components will decrease H2O/Ce. If variations in H2O/Ce correlate with radiogenic isotopic compositions, then the hydration or dehydration event probably occurred at the same time as the fractionation between parent and daughter elements.
Water in plume-influenced MORB from the South Atlantic
Figure 1 Comparison of radiogenic isotopic compositions, H2O, K2O and H2O/Ce in basaltic glasses. Data are from the Discovery and Shona anomalies9 along the southern Mid-Atlantic Ridge, and from the Easter microplate and Easter–Salas y Gomez seamount chain (EMP-ESC) in the southeast Pacific35,36. a, 87Sr/86Sr versus 206Pb/204Pb. Isotopic compositions for mantle end-members are from ref. 51. Open symbols (Discovery and Shona group 2) represent samples having an EM-type mantle plume component. Filled symbols (Shona group 1 and EMP-ESC) represent samples containing a FOZO-type or mildly-HIMU plume component. b, H2O versus K2O. Water concentrations increase with increasing K2O concentrations; however, EM-type basalts contain less water at a given K2O than C-type basalts. c, H2O/Ce versus 87Sr/86Sr for the South Atlantic and Easter microplate. EM-type basalts have significantly lower H2O/Ce than FOZO-type basalts. Several samples have high H2O/Ce ratios (.300) that do not correlate with major- or trace-element or isotopic compositions. These samples may have assimilated water during transport and eruption. Symbols are as in b. 386
South Atlantic samples were dredged along the Mid-Atlantic Ridge between 408 S and 548 S during the RV Maurice Ewing EW93-09 cruise (see Supplementary Table S1). This region contains two distinct geochemical, gravity and bathymetric anomalies. The northern ‘Discovery’ anomaly is centred at 47.58 S and is influenced by the Discovery plume9,33. The southern ‘Shona’ anomaly is centred at 51.58 S (refs 9, 33, 34). We compare the South Atlantic samples to EMP and ESC glasses because both areas involve plume–ridge interaction, and Pacific and South Atlantic MORB have similar regional H2O/Ce ratios23, and thus similar depleted end-member compositions. Radiogenic isotopic data for EMP-ESC basalts (Fig. 1a) have been interpreted as the result of binary mixing between depleted MORB and mildlyHIMU FOZO (Salas y Gomez plume) sources35,36. Isotopic compositions of Discovery glasses show a wide range of 87 Sr/86Sr at a relatively constant 206Pb/204Pb, consistent with pseudo-binary mixing of DMM, FOZO and an EM component, specifically LOMU9 (Fig. 1a). Contribution of this EM component is least for samples 26, 28, 34 and 40, and greatest for sample 7, 9, 33 and 25 (ref. 37)37. Whether this component is recycled in the upper or lower mantle is not known. However, the nearly solar 21Ne/22Ne ratios and high 3He/4He in some of the Discovery MORB glasses (samples 2–5) suggest recycling at least to the 660-km discontinuity, where more primitive mantle could have been entrained by the Discovery plume38. Shona glasses form two groups in Sr–Nd–Pb–Hf isotopic space. Shona group 1 samples (dredges 11, 14, 15, 17 and 19) show mixing between DMM and FOZO with mild HIMU influence, similar to the Easter trend. Shona group 2 samples (dredges 18, 20, 21, 22 and 23) show additional influence of an EM component (EM1-EM2) thought to represent recycled oceanic crust þ pelagic sediments15. Water correlates positively with K2O for the South Atlantic, but with very different slopes defined by Shona and Discovery samples (Fig. 1b). Linear correlations between H2O and K2O for Shona group 1, Easter, and other South Atlantic (31-468S) MORB samples are indistinguishable (Fig. 1b). One outlier (sample 19) has more water at a given K2O (about 0.1 wt% extra) than samples forming the main trend, with H2O/Ce similar to several ESC samples, possibly related to serpentine assimilation during ascent23. For the remaining samples, water is more compatible than K and less compatible than Ce, consistent with the behaviour of these elements in magmatic systems. Thus, the characteristics of the plume and DMM end-members for the Shona group 1 samples are similar to those of the EMP-ESC30, for which the FOZO component has H2O/ K2O ¼ ,0.8 and H2O/Ce ¼ 210 ^ 30 and the DMM component has H2O/K2O ¼ ,5 and H2O/Ce ¼ 150 ^ 10 (Fig. 1c). Shona group 2 and Discovery samples (87Sr/86Sr . 0.703) that contain an EM component have distinctly lower H2O concentrations and lower H2O/Ce than Shona group 1 and Easter samples. Shona group 2 data form trends sub-parallel to Shona group 1 data but at progressively lower H2O content and 206Pb/204Pb. The
© 2002 Nature Publishing Group
NATURE | VOL 420 | 28 NOVEMBER 2002 | www.nature.com/nature
articles Discovery samples form a distinct binary mixing line converging with that of the Easter group at the depleted source end (Fig. 1b, c). Discovery samples with the most definitive EM signature have H2O/ K2O half that of the Easter samples. The observation that sample 25 has relatively low water concentration and H2O/Ce is very important. Even though it has the highest 87Sr/86Sr ratio, reflecting a mantle source with time-integrated enrichment of incompatible elements (for example, high Rb/Sr), it is not the most enriched in incompatible elements, perhaps related to a recent episode of
melting and melt extraction39. Sample 25 has only 0.15 wt% water. In order for this sample to have degassed significant quantities of water, it would have to have been erupted at or below the water solubility pressure of ,3 bar (30 m water depth), much shallower than its actual eruption depth at 2,032 m. It is almost certain that its low water concentration is a primary magmatic characteristic. Samples most enriched in 87Sr/86Sr (7, 25 and 33) have the lowest H2O/ K2O. If these ratios were caused by the same type of (magmatic) process as observed for the Easter samples, then these samples should also have the highest H2O/Ce. Instead, samples with the lowest H2O/K2O have low H2O/Ce, suggesting control by a different process that selectively lowers water content. Thus, the EM components in the mantle sources for the Shona and Discovery samples are associated with lower water concentrations and lower ratios of water to similarly incompatible elements compared to the common plume component (FOZO).
Water in plume-influenced MORB from the North Atlantic
Figure 2 Comparison of radiogenic isotopic compositions, H2O, K2O and H2O/Ce in basaltic glasses from the North Atlantic. a, 87Sr/86Sr versus 206Pb/204Pb. Azores platform basalts43 (black filled triangles) are consistent with mixing between DMM and C-type mantle components. Great Meteor anomaly basalts (blue open circles) are consistent mixing between EM-type and FOZO-type mantle components. b, H2O versus K2O. K2O data from ref. 40. Measured dissolved H2O concentrations are shown as black half-filled squares (Azores platform) and blue crossed squares (Great Meteor). Calculated bulk H2O concentrations (see Supplementary Tables S1 and S2) are shown as black filled triangles (Azores platform) and blue open circles (Great Meteor). Lines connect dissolved and bulk water concentrations for individual samples. Calculated bulk H2O concentrations for dredge 17 from the summit of the Azores platform are probably low owing to open-system degassing (J.E.D., manuscript in preparation). EM-type basalts have less water at a given K2O content than FOZO-type basalts from the North Atlantic. c, H2O/Ce versus 87Sr/86Sr for the North Atlantic. EM-type basalts have lower H2O/Ce than FOZO-type basalts and depleted MORB. Symbols as in b. NATURE | VOL 420 | 28 NOVEMBER 2002 | www.nature.com/nature
The north Mid-Atlantic Ridge samples were dredged by the RV Atlantis II (FAZAR expedition40) between 33.28 and 40.58 N with depths varying between 926 and 3,900 m (see Supplementary Table S2). Two distinct geochemical anomalies occur within this region, including the long-wavelength anomaly associated with the intersection of the Azores archipelago with the Mid-Atlantic Ridge at ,38–398 N (Azores platform) and a short-wavelength anomaly centred at 358 N (the Great Meteor anomaly41, used here, or the ‘north of Oceanographer fracture zone anomaly’42). Sr and Pb isotopic data43 (Fig. 2a) show that Azores platform basalts form a linear mixing array between DMM and FOZO. In contrast, basalts from the Great Meteor anomaly form a mixing array between FOZO and an EM component, with samples 9-10, 45-3 and 44-1 having the highest 87Sr/86Sr ratios. An exception is 49-3, which falls within the Azores platform array. Because of the high vesicularity of many of the Azores platform and Great Meteor samples, loss of water into vesicles may have been significant. Both measured dissolved and calculated bulk water concentrations (see Supplementary Information) are shown in Fig. 2b and c. Samples from dredges 17, 19, 21 and 22 from the summit of the Azores platform (926–1,950 m) are highly vesicular (25–60 vol.%) and have lost about 10–30% of their initial water into vesicles. Dissolved water concentrations in the less vesicular Azores platform samples correlate positively with K2O and have H2O/Ce ratios (253 ^ 33) consistent with previously reported values for the North Atlantic28. Estimated bulk H2O/Ce for the highly vesicular samples (247 ^ 54) are similar to values for nonvesicular samples for all but the shallowest dredge (17), which may have suffered some gas loss (open-system degassing). Great Meteor anomaly samples are also highly vesicular (5– 60 vol.%), with vesicularity correlating roughly inversely with depth (5 vol.% at 3,900 m to 60 vol.% at 1,657 m). Samples with the largest EM component have dissolved and bulk water concentrations at a given K2O (Fig. 2b) that are significantly lower than undegassed samples from the Azores platform. H2O/Ce is also 30% lower than the mean for the Azores platform samples (168 ^ 29 versus 253 ^ 33). Though the correlation is rough, extrapolation from FOZO through the Great Meteor data yields H2O/Ce , 100 at 87 Sr/86Sr . 0.705. Thus, like the Discovery samples, EM components in the North Atlantic are depleted in water relative to FOZO.
Discussion In each setting described above, there are two distinct mantle plume components: one close to FOZO (common mantle plume component), and the other enriched in recycled material of continental derivation (EM). In each case, the ratio of water to other similarly incompatible elements is lower in the EM component than in FOZO. Our data show that basalts influenced by EM mantle
© 2002 Nature Publishing Group
387
articles components are depleted in water (‘damp’ plumes) relative to those influenced by FOZO (‘wet’ plumes). What is FOZO?
We have shown that FOZO plumes are ‘wetter’ than EM plumes, but we still need to address the origin of H2O in these mantle components. Though the origin of FOZO is an area of active debate, the following characteristics are generally accepted. First, FOZO is not pristine primitive mantle. Paradoxically, Sr- and Nd-isotopic compositions require moderate time-integrated depletion in incompatible elements (Rb/Sr lower and Sm/Nd higher than primitive mantle), whereas the present-day source is enriched in incompatible elements6,44. In addition, Pb-, Os- and O-isotopic compositions and trace-element ratios (Nb/U, Nb/Th and Ce/Pb) suggest involvement of a HIMU component6,10,21–25,44. Second, high concentrations of Ni, Cr and heavy rare-earth elements in FOZO require melting of a peridotitic source, and are inconsistent with melting of a purely eclogitic (HIMU) source44,45. Third, FOZO typically has elevated 3He/4He, indicating involvement of mantle that is less degassed and more primitive than MORB mantle, probably located deeper than the MORB source17,18 . High 3 He/4He also rules out generation of FOZO by mixing between DMM and HIMU components45. Lastly, the globally widespread occurrence of FOZO in ocean island basalts and carbonatites makes it unlikely that FOZO is the product of melting heterogeneous recycled crustal materials that make up only 4–10% of mantle volume18,45. Similar to a recent model by Donnelly et al.44, we suggest that FOZO represents dominantly primitive or slightly depleted peridotitic mantle (,97–99%), metasomatized by a small amount (1–3%) of a small extent melt (2–6%) of a HIMU component. If the lower mantle is composed of peridotite with ‘streaks’ of recycled eclogitized oceanic crust (HIMU), then this metasomatism could occur during ascent of plumes, a process recently proposed to explain major-element heterogeneity in picrites and komatiites46. In such a mixture, concentrations of noble gases and compatible elements (Ni, Cr and heavy rare-earth elements) in the FOZO source come almost completely from the peridotitic mantle, whereas roughly half to two-thirds of the highly incompatible elements are supplied by the metasomatic HIMUmelt. Thus, water in FOZO plumes is a combination of primitive and recycled water. H2O in mantle sources of FOZO and HIMU plumes
On the basis of eclogite measurements47, subducted oceanic crust
Table 1 Dehydration efficiency Rock
H2O/Ce
H2O (wt%)
250 150 2,500–5,000 1,280
0.3 0.2 2–3 7.3
Ce (p.p.m.)‡
Dehydration (%)
.............................................................................................................................................................................
Crustal components: Pre-subduction Atlantic MORB* 12 Pacific MORB* 14 Mature oceanic crust* ,6 Global subducted sediment† 57 Crustal components: Post-subduction Mature oceanic crust ,6 Global subducted sediment 57 Mantle components DMM* 0.5 Atlantic FOZOk 3 Pacific FOZO{ 3.8 EM# 4
,100 ,100
0.06§ ,0.57
200 250 200 ,100
0.010 0.075 0.075 ,0.04
97 .92
............................................................................................................................................................................. * GERM (Geochemical Earth Reference Model) database (http://earthref.org/databases/GERMRD/ main.htm). † GLOSS (Global Subducting Sediment)48. ‡ Represents conservative estimate—some rare-earth elements are lost from subducting lithosphere during subduction. § Bulk water concentration in eclogite47. kAsimow, P., C.L. and J.E.D., manuscript in preparation. {From refs 27 and 30. # EM sources have Ce contents similar to HIMU sources6 and slightly greater than FOZO-type sources. EM mantle sources have roughly half as much water as FOZO-type sources (this study).
388
(HIMU source) is low in water, with ,600 p.p.m. H2O, ,6 p.p.m. Ce, and H2O/Ce ¼ ,100. These data suggest that subducted oceanic crust is efficiently (,97%) dehydrated during subduction (Table 1). In contrast, mantle sources of FOZO plumes have ,750 p.p.m. H2O and H2O/Ce of ,210 in the Pacific and South Atlantic and ,250 in the North Atlantic (ref. 25, and P. Asimow, C.L. and J.E.D., manuscript in preparation). The high H2O/Ce in FOZO plumes cannot be derived from recycled oceanic crust; a significant amount of water must be juvenile, left over from planetary accretion. We believe that it is the presence of this juvenile water that makes FOZO plumes ‘wet’. Water in EM components
We have shown that recycled oceanic crust is efficiently dehydrated during the subduction process. The lower H2O/Ce of EM plumes relative to FOZO plumes indicates that recycled continentally derived components are also efficiently dehydrated. For example, an estimate of globally subducted sediment48 contains 57 p.p.m. Ce and ,7.3 wt% H2O (H2O/Ce ¼ 1,280). Reduction of the H2O/Ce to ,100, lower than the ambient upper mantle of .150, requires .92% dehydration of the water from sediments (Table 1). These dehydrated sediments are still fairly hydrous (,0.57 wt% H2O), but addition of only a few per cent sediments to HIMU will only increase the water concentration by 50 to 150 p.p.m. Thus our data suggest that dehydration is efficient for all recycled materials. These estimates of dehydration are consistent with other approaches that have considered mass balance at the subduction zone itself49. Discovery of dense hydrous magnesian silicate phases in laboratory experiments is generally considered consistent with a scheme whereby water may be efficiently transported into the deep mantle during subduction50. However, laboratory generation of phases capable of holding significant amounts of water does not constitute evidence that these phases in fact carry significant water to the deep mantle. Our data suggest that much of the recycled material that has passed through the subduction ‘factory’ has H2O/Ce , 100. Thus, as is the case for Pb (refs 21–23), the subduction process effectively partitions water into the exosphere (mantle wedge, crust, ocean and atmosphere), resulting in time-integrated depletion of water relative to other incompatible elements in the recycled lithosphere and accompanying continental-derived materials. Thus, like a wellsqueezed sponge, the residual slab and EM plume components are not ‘wet’, but ‘damp’. A Received 28 August; accepted 9 October 2002; doi:10.1038/nature01215. 1. Hirth, G. & Kohlstedt, D. L. Water in the oceanic upper mantle: implications for rheology, melt extraction and the evolution of the lithosphere. Earth Planet. Sci. Lett. 144, 93–108 (1996). 2. Hirose, K. & Kawamoto, T. Hydrous partial melting of lherzolite at 1 GPa; the effect of H2O on the genesis of basaltic magmas. Earth Planet. Sci. Lett. 133, 463–473 (1995). 3. Abe, Y., Ohtani, E., Okuchi, T., Righter, K. & Drake, M. Origin of the Earth and Moon (eds Canup, R. M. & Righter, K.) 413–433 (Univ. Arizona Press, Tucson, 2000). 4. Schmidt, M. W. & Poli, S. Experimental base water budgets for dehydrating slabs and consequences for arc magma generation. Earth Planet. Sci. Lett. 163, 361–379 (1998). 5. Schilling, J.-G., Bergeron, M. B. & Evans, R. Halogens in the mantle beneath the North Atlantic. Phil. Trans. R. Soc. Lond. A 297, 147–178 (1980). 6. Hofmann, A. W. Mantle geochemistry: the message from oceanic volcanism. Nature 385, 219–229 (1997). 7. White, W. M. Sources of oceanic basalts: Radiogenic isotopic evidence. Geology 13, 115–118 (1985). 8. Zindler, A. & Hart, S. Chemical geodynamics. Annu. Rev. Earth Planet. Sci. 14, 493–571 (1986). 9. Douglass, J., Schilling, J.-G. & Fontignie, D. Plume-ridge interactions of the Discovery and Shona mantle plumes with the southern Mid-Atlantic Ridge (408-558S). J. Geophys. Res. 104, 2941–2962 (1999). 10. Weaver, B. L. The origin of ocean island basalt end-member compositions: Trace element and isotopic constraints. Earth Planet. Sci. Lett. 204, 381–397 (1991). 11. Chauvel, C., Hofmann, A. W. & Vidal, P. HIMU-EM: the French Polynesian connection. Earth Planet. Sci. Lett. 110, 99–119 (1992). 12. Blichert-Toft, J., Frey, F. A. & Albare`de, F. Hf isotope evidence for pelagic sediments in the source of Hawaiian basalts. Science 285, 879–882 (1999). 13. McKenzie, D. & O’Nions, R. K. Mantle reservoirs and ocean island basalts. Nature 301, 229–231 (1983). 14. Kamenetsky, V. S. et al. Remnants of Gondwanan continental lithosphere in oceanic upper mantle: Evidence from the South Atlantic Ridge. Geology 29, 243–246 (2001). 15. Andres, M., Blichert-Toft, J. & Schilling, J.-G. Hafnium isotopes in basalts from the Southern MidAtlantic Ridge from 408S to 558S: Trace element evidence for contamination of ambient asthenospheric mantle. Earth Planet. Sci. Lett. (in the press).
© 2002 Nature Publishing Group
NATURE | VOL 420 | 28 NOVEMBER 2002 | www.nature.com/nature
articles 16. Hanan, B. B. & Graham, D. W. Lead and helium isotope evidence from oceanic basalts for a common deep source of mantle plumes. Science 272, 991–995 (1996). 17. Hart, S. R., Hauri, E. H., Oschmann, L. A. & Whitehead, J. A. Mantle plumes and entrainment: The isotopic evidence. Science 256, 517–520 (1992). 18. Hauri, E. H., Whitehead, J. A. & Hart, S. R. Fluid dynamic and geochemical aspects of entrainment in mantle plumes. J. Geophys. Res. 99, 24275–24300 (1994). 19. Alle`gre, C. J., Hofmann, A. W. & O’Nions, R. K. The argon constraints on mantle structure. Geophys. Res. Lett. 23, 3555–3557 (1996). 20. Hofmann, A. W. Nb in Hawaiian magma: constraints on source composition and evolution. Chem. Geol. 57, 17–30 (1986). 21. Hofmann, A. W., Jochum, K. P., Seufert, M. & White, W. M. Nb and Pb in oceanic basalts: new constraints on mantle evolution. Earth Planet. Sci. Lett. 79, 33–45 (1986). 22. Miller, D. M., Goldstein, S. L. & Langmuir, C. H. Cerium/lead and lead isotope ratios in arc magmas and the enrichment of lead in the continents. Nature 368, 514–520 (1994). 23. Chauvel, C., Goldstein, S. L. & Hofmann, A. W. Hydration and dehydration of oceanic crust controls Pb evolution in the mantle. Chem. Geol. 126, 65–75 (1995). 24. Shirey, S. B. & Walker, R. J. The Re-Os isotope system in cosmochemistry and high-temperature geochemistry. Annu. Rev. Earth Planet. Sci. 26, 423–500 (1998). 25. Eiler, J. M. Stable Isotopes in Igneous Processes (eds Cole, D. and Valley, J. W.) (Mineralological Society of America, Washington DC, 2002). 26. Menzies, M. & Murthy, V. R. Enriched mantle: Nd and Sr isotopes in diopsides from kimberlite nodules. Nature 283, 634–636 (1980). 27. Dixon, J. E. & Clague, D. A. Volatiles in basaltic glasses from Loihi Seamount, Hawaii: Evidence for a relatively dry plume component. J. Petrol. 42, 627–654 (2001). 28. Michael, P. J. Regionally distinctive sources of depleted MORB: Evidence from trace elements and H2O. Earth Planet. Sci. Lett. 131, 301–320 (1995). 29. Danyuschevsky, L. V., Eggins, S. M., Falloon, T. J. & Christie, D. M. H2O abundance in depleted to moderately enriched mid-ocean ridge magmas; Part I: Incompatible behaviour, implications for mantle storage, and origin of regional variations. J. Petrol. 41, 1329–1364 (2000). 30. Simons, K., Dixon, J. E., Schilling, J.-G., Kingsley, R. & Poreda, R. Volatiles in basaltic glasses from the Easter-Salas y Gomez Seamount Chain and Easter Microplate: Implications for geochemical cycling of volatile elements. Geochem. Geophys. Geosyst. 3, doi: 10.1029/2001GC000173 (2002). 31. Dixon, J. E., Clague, D. A., Wallace, P. & Poreda, R. Volatiles in alkalic basalts from the North Arch Volcanic Field, Hawaii: Extensive degassing of deep submarine-erupted alkalic series lavas. J. Petrol. 38, 911–939 (1997). 32. Michael, P. J. & Cornell, W. C. Influence of spreading rate and magma supply on crystallization and assimilation beneath mid-ocean ridges: Evidence from chlorine and major element chemistry of midocean ridge basalts. J. Geophys. Res. 103, 18325–18356 (1998). 33. Douglass, J., Schilling, J.-G. & Kingsley, R. H. Influence of the Discovery and Shona mantle plumes on the southern Mid-Atlantic Ridge: Rare earth evidence. Geophys. Res. Lett. 22, 2893–2896 (1995). 34. le Roex, A. P., Dick, H. J. B., Gulen, L., Reid, A. M. & Erlank, A. J. Local and regional heterogeneity in MORB from the Mid-Atlantic Ridge between 54.58S and 518S: Evidence for geochemical enrichment. Geochim. Cosmochim. Acta 51, 541–555 (1987). 35. Kingsley, R. H. & Schilling, J.-G. Plume-ridge interaction in the Easter-Salas y Gomez seamount chain-Easter Microplate system: Pb isotope evidence. J. Geophys. Res. 103, 24159–24177 (1998). 36. Cheng, Q. C., Macdougall, J. D. & Zhu, P. Isotopic constraints on the Easter Seamount Chain source. Contrib. Mineral. Petrol. 135, 225–233 (1999).
NATURE | VOL 420 | 28 NOVEMBER 2002 | www.nature.com/nature
37. Douglass, J. & Schilling, J.-G. Systematics of three-component, pseudo-binary mixing lines in 2D isotope ratio space representations and implications for mantle plume-ridge interaction. Chem. Geol. 163, 1–23 (2000). 38. Sarda, P., Moreira, M., Staudacher, T., Schilling, J.-G. & Alle`gre, C. J. Rare gas systematics on the southernmost Mid-Atlantic Ridge: Constraints on the lower mantle and the Dupal source. J. Geophys. Res. 105, 5973–5996 (2000). 39. le Roux, P. J. et al. Mantle heterogeneity beneath the southern Mid-Atlantic Ridge: trace element evidence for contamination of ambient asthenospheric mantle. Earth Planet. Sci. Lett. (in the press). 40. Langmuir, C. H., Klinkhammer, G. & Bougalult, H. Cruise Report: Evaluation of the Relationships among Segmentation, Hydrothermal Activity and Petrological Diversity on the Mid-Atlantic Ridge (FAZAR) Technical Report LDEO-92-3 (Lamont-Doherty Earth Observatory, Columbia Univ., Palisades, 1992). 41. Rideout, M. L. & Schilling, J.-G. Rare-earth elements, 87Sr/86Sr, and 143Nd/144Nd mantle source variations. Init. Rep. DSDP 82, 483–496 (1985). 42. Shirey, S. B., Bender, J. F. & Langmuir, C. H. Three-component isotopic heterogeneity near the Oceanographer transform, Mid-Atlantic Ridge. Nature 325, 217–223 (1987). 43. Dosso, L. et al. The age and distribution of mantle heterogeneity along the Mid-Atlantic Ridge (31418N). Earth Planet. Sci. Lett. 170, 269–286 (1999). 44. Donnelly, K., Langmuir, C. H., Goldstein, S. L. & Lagatta, A. The origin of alkali and ocean island basalts: contradictions and solutions. Eos 82, F1402 (2001). 45. Bell, K. & Tilton, G. R. Probing the mantle: The story from carbonatites. Eos 83, 273–277 (2002). 46. Gibson, S. A. Major element heterogeneity in Archean to Recent mantle plume starting-heads. Earth Planet. Sci. Lett. 195, 59–74 (2002). 47. Nadeau, S., Philippot, P. & Pineau, F. Fluid inclusion and mineral isotopic compositions (H-C-O) in eclogitic rocks as tracers of local fluid migration during high-pressure metamorphism. Earth Planet. Sci. Lett. 114, 431–448 (1993). 48. Plank, T. & Langmuir, C. H. The chemical composition of subducting sediment and its consequences for the crust and mantle. Chem. Geol. 145, 325–394 (1998). 49. Peacock, S. M. Fluid processes in subduction zones. Science 248, 329–337 (1990). 50. Thompson, A. B. Water in the Earth’s upper mantle. Nature 358, 295–302 (1992). 51. Schilling, J.-G. Les Isotopes Radioge´niques en Ge´ologie: Chronologie Geochimie (ed. Duthou, J. L.) 1–34 (Socie´te´ Franc¸aise de Mine´ralogie et de Cristallographie, Paris, 1992).
Supplementary Information accompanies the paper on Nature’s website (ç http://www.nature.com/nature). Acknowledgements This research was funded by the NSF. A review by N. Arndt improved the manuscript. Competing interests statement The authors declare that they have no competing financial interests. Correspondence and requests for materials should be addressed to J.E.D. (e-mail: [email protected]).
© 2002 Nature Publishing Group
389
